An academic base for neural-network research
Founded in 1827, the University of Toronto became Geoffrey Hinton's academic home in 1987. Its account of his 2024 Nobel Prize identifies him as professor emeritus and describes the university research behind the 2012 AlexNet result. The award was shared by Hinton and John Hopfield, rather than awarded to the university as an institution. [1][2]
The 2006 deep belief nets paper by Hinton, Simon Osindero and Yee Whye Teh describes greedy layer-wise learning followed by fine-tuning. It provided a training procedure for the deep generative networks studied in that paper. [4][3]
AlexNet used a deep convolutional network and an efficient GPU implementation for ImageNet classification. The 2012 paper reports the architecture and benchmark experiment, connecting Toronto's neural-network research to a measurable computer-vision result. Its historical scores are not present-day performance comparisons. [5]